Other ways of understanding Cohen’s d
Cohen’s d (d) is quite a used measure of the size of the effect and its report is compulsory necessary in sta-tistical analyzes. Nevertheless, researchers report that the difference between two distributions is small (d > .20). However, the interpretation of this coefficient is not clear in p...
Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Artículo revista |
| Lenguaje: | Español |
| Publicado: |
Instituto de Investigaciones Psicológicas (IIPSI, Conicet-UNC)
2018
|
| Materias: | |
| Acceso en línea: | https://revistas.unc.edu.ar/index.php/revaluar/article/view/22305 |
| Aporte de: |
| Sumario: | Cohen’s d (d) is quite a used measure of the size of the effect and its report is compulsory necessary in sta-tistical analyzes. Nevertheless, researchers report that the difference between two distributions is small (d > .20). However, the interpretation of this coefficient is not clear in psychology studies. In this sense, it is necessary to con-vert the d into a probability measure to facilitate the inter-pretation of the distributions that are object of comparison. Among the most frequent measures are: Cohen’s U3, the superposition coefficient (OVL), the probability of superi-ority (PS) and the number needed to treat (NNT), which can be considered as alternative measures of the magnitude of a difference. For such purposes, R codes that can be easi-ly used by the researchers are provided, as well as a table showing the modifications of the alternative measures be-fore the increase in the size of the effect. |
|---|